US grant giving agencies have cut overhead reimbursements to 15% from 60% earlier. Without the overhead, basic research is not possible. How should universities, donors and politicians respond?
US grant giving agencies have cut overhead reimbursements to 15% from 60% earlier. Without the overhead, basic research is not possible. How should universities, donors and politicians respond?
One of the seismic shocks to the business model of a university is the new administration’s decision, in February 2025, to limit the reimbursement of indirect costs to 15% of the value of the grant. Those rates, before February 2025, used to be about 50% – 60%. At a very simplistic level, for a $1 million grant, approximately 60% went to the university for recovery of administrative overhead and the other 40% was available for the actual research (its more complicated than that, as explained later). These overhead cuts have been implemented at the NIH (National Institute of Health), DOE (Department of Energy), NSF (National Science Foundation) and other such federal agencies.
What makes things worse is that the 15% overhead rate applies to multi-year grants promised in the past as well as new grants moving forward. This means universities that have counted on a 60% recovery rate will only get 15%, leaving a hole of around 45% in their future cash inflows. Part of the motivation appears to be that grants by private donors such as the Gates Foundation or the Chan-Zuckerberg initiative only reimburse 10-15% of the grant value towards overhead.
Because I work at a business school, I have never raised a grant from a government agency for my research. I have some experience with grants from private donors. I asked four knowledgeable individuals for their take on this development: (i) a retired CEO of a health services publicly listed firm (henceforth the “CEO”); (ii) a business school academic with extensive experience in analyzing healthcare and energy companies (“the academic”); (iii) an advisor to large university endowments (“the advisor”); and (iv) a well-known academic doctor who works for a university hospital with a track record of inventing life-saving medical devices (“the doctor”). Here are edited excerpts from our conversations.
THE PROBLEM
The actual overhead reimbursement process is complicated
The doctor states that the actual process is much more complicated than how the lay public and even the informed commentators think of the problem. The doctor puts in a budget of $500,000 to conduct the research. The university gets a maximum of say 60% as overhead reimbursement. So, the maximum amount that the NIH gives is $800,000 ($500,000+60%). The actual amount the NIH pays depends on how the budget is spent. Say $100,000 of the $500,000 is used for capital expenditure, for instance, to buy a centrifuge. The university will not get reimbursed for overhead corresponding to the capex spend. The actual reimbursement rate collected by the universities is around 35% on average. How?
It turns out that reimbursement from NIH is triggered only for specific accounts for which the grant money is spent and the rates differ depending on the nature of the expense. For instance, reimbursement for administrative costs is capped at something like 26%, fellowship and training grant costs are limited to 8%, and equipment and tuition remission costs are capped at 0%, meaning the University gets no reimbursement of overhead from the NIH on equipment such as the centrifuge purchase and tuition payments of grad students working on the research.
This is an important aspect of the process that the public potentially does not know about (I did not). Think of the 60 plus percentage rates advertised in the press as the ceiling or the maximum reimbursement rates. The actual effective rates of recovery from NIH are closer to 35% or so. This confusion is apparent in the comment I got from the CEO (the following point).
Capex vs Opex problem? Yes and no
The CEO points out that ultimately the business question boils down to recovery of some of the capex (capital expenditure) via revenues or grants. “in the previous regime, the 60% overhead rate was used partly to recover the fixed costs associated with running the facility such as depreciation and capital improvements. Now, most of such capital costs will not be recovered from the federal agency.” The CEO goes on to suggest that he could not run a successful publicly listed company with these economics. That is, maintenance capex, covered by indirect costs, cannot be 60% of revenue of a private sector enterprise. The 15% rate seems closer to what we would see in the private sector.
Of course, the CEO has conflated capex with opex (operating expenses) overhead that is common to all projects. “Indirect costs” in the NIH context also include operating expenses that are common to many projects such as utility bills, information technology and administrative staff expenses. So, the old 60% rate does not strictly go only to recover capex spending. On top of that, as pointed out by the doctor, the effective rate of reimbursement is around 35%, not 60%.
Overhead rates in the private sector
To empirically examine the CEO’s claims that private businesses are more efficient with overhead, I looked at the “overhead” rate of the top three pharmaceutical companies, Pfizer, Johnson and Johnson and Roche, with sales of around $55-$65 billion each in 2024. Any comparison is difficult as the purpose of a modern pharmaceutical company is more advertising and getting FDA approvals rather than basic research, as in a university. Even then, the data are interesting.
Pfizer reports sales of $64 billion in 2024. In its expenses column, Pfizer reports cost of sales ($18 billion), selling, information and administrative expenses ($15 billion), R&D ($11 billion), amortization of intangible assets ($5 billion), restructuring ($2 billion) and other expenses ($4.4 billion). All classifications are subject to some measurement error but not enough to be crippling. But the closest estimate of “indirect” costs for Pfizer constitutes all expenses, other than cost of sales and R&D. That works out to an “overhead” rate of 41% ($15+5+2+4.4/$64 billion). I have left out taxes to ensure pre-tax comparisons as universities are generally exempt from income taxes.
Let’s repeat the analysis for Johnson and Johnson. J&J reports sales of $88 billion. I have assumed that “overhead” or indirect costs for J&J include SGA ($23 billion), other expenses ($4.7 billion). That gives us an overhead rate of 31% ((23+4.7)/88.
Roche reported sales of 62 billion CHF in 2024. SGA (15 billion CHF), other expenses (2.5 billion CHF) and interest expense (1.6 billion CHF) cumulatively amount to 31% “overhead” rate.
Three important takeaways emerge from this analysis:
- Without the “overhead,” there is no way to run a Pfizer, Roche or a J&J. The same point applies to university research. That is, there is no university research without overhead.
- The overhead rates in the private sector are not that much better relative to the true effective overhead rates charged in universities.
- The overhead rates in the private sector do not account for the “market value” of the basic research that they get from universities. If university basic research were to disappear, someone else will have to conduct such research at “market prices” that are far higher than taxpayer supported basic research at universities.
The CEO’s response
When I showed the data for Pfizer, J&J and Roche to the CEO, he admitted that there can be considerable bloat even in the private sector at large companies with lots of employees. One blockbuster drug pays for all the failures and administrative bloat in private pharma.
Unfortunately, as I have suggested before, universities suffer a huge reporting problem: the costs of basic research are recorded in the books of the university whereas the benefits are accrued by private drugmakers and other companies. Hence, it becomes easier to blame universities for high costs and give the private sector a pass as long as listed firms report returns above cost of capital.
Markups for service businesses
I counter-argued, with both the CEO and the advisor, that service businesses such as consultants and law firms routinely build in large mark-ups when they bill their clients. Universities operate, at best, with a no-profit motive and want to simply recover the indirect costs of operations. The CEO countered, “a business with a cost-plus billing arrangement never wants to cut costs.” We can agree that there is definitely some wastage and bloat in the cost structure of research facilities. But is the bloat as high as 45% (60-15) of the contract value? Most likely not.
NIH and other grants are very competitive
The doctor, who happens to be a very distinguished inventor, points out that NIH grants are very competitive and his success rate at securing a grant is barely 10%. Chinese doctors, with perhaps half of his stature in the field, get funded automatically for 100% of their grant proposals from the Chinese government. Hence cutting funding to grants that are subject to rigorous competition highlights the enormity of the problem that US science faces at this moment, relative to its geopolitical rival, China.
OTHER NUANCES ON THE OVERHEAD QUESTION
15% new overhead rate across the board average masks important nuances
The academic points out that things are more complicated than imposing a uniform rate of 15%: “healthcare researchers exist in an institution that is broader than that found in any corporate entity. Even those within large corporate sub-units do not have to contribute much to the corporate overhead. Also, it is hard to know how much is really for the lab/capex when looking at averages.
Consider cancer treatments for a minute. Clearly the machinery to execute radiation or laser treatments is capex but what about the hospital facilities for the chemo or hormone treatments which are being researched? I am also not convinced that there is a lot of capex outside the buildings in most medical research. I would expect this more in Physics and Engineering.”
These comments suggest a more bespoke response to each cut. That is, the response to an NIH cut might differ from that of an NSF or a DOE cut.
What about the overhead at Federal Agencies themselves?
I was curious about the indirect costs of the NIH itself. I could not find a detailed set of financial statements. The summary statements state that NIH spent $4.4 billion in wages, salary and benefits for around 21,000 employees in fiscal year 2024. The CEO states that an NIH grant can take 9-12 months to process. Does that suggest bloat or understaffing at the NIH? The new administration has slashed the NIH’s budget from $45 billion to $27 billion. Will these slashed budgets simply increase NIH review times for proposals from the current 9-12 months with consequent delays in deploying research funds for discovering new drugs and treatments?
What about overhead at FFRDCs?
The advisor pointed me to the budgets for FFRDCs (the Federally Funded Research and Development Centers) such as the CMS Alliance to Modernize Healthcare or the Frederick National Laboratory for Cancer Research. Are these agencies more efficient than universities at commissioning research? I looked around but did not find adequate data to answer these questions. The advisor’s gut feel is that universities are more efficient at getting research done.
This leads to the question of what universities should do?
STRATEGY
Prioritize
This is perhaps obvious, but the CEO suggests prioritizing projects that will likely have a large impact on future outcomes and sacrifice low impact, low priority projects. I asked what high impact might look like. That is likely to be specific to individual healthcare facilities or universities, but he suggests AI and work on life saving diseases.
Unfortunately, universities might start focusing less on clinical work or work that involves lot of set up costs and capex. The focus might shift to computational work of the sort I do for instance, with accounting and finance research, where we crunch large datasets sitting in our offices. I don’t need a lab or expensive hardware. Of course, the work done in labs might be more valuable to society than accounting or finance research but that is a different topic. But the axe is likely to fall on staff and administrators than on research personnel, all else constant.
Because all money is fungible, resources to support the research will have to come from one of the other revenue sources of the university such as tuition, or by cutting costs via fewer scholarships or reduced support to graduate students or smaller number of faculty hired. Ironically, more of the administrative burden addressed for now by assistants will fall on researchers, whose opportunity cost of time is greater than that of assistants! All this points to a decline in the research footprint of American universities in the future.
“Fund raise” for capex projects and indirect costs
Deans and presidents routinely raise funds for new buildings on campus. Do we need to identify private donors to fund endowments that will cover facilities and “indirect” costs? Of course, this is easier said than done, as donors have strong feelings about not wasting their funds on “overhead.”
Perhaps the funding shock is a chance for deans and presidents to educate donors on the need to cover both “direct” and “indirect” costs to cover the expenses associated with conducting research and hence change the model of giving to a research university. Perhaps we need new language and stop using the word, “overhead.” How about “management costs” or “operating costs” or something along those lines?
The IP (intellectual property) question
The advisor goes to on ask, “has anyone recently criticized the fairness to the taxpayer of IP developed with government grants remaining with the grantee?” The doctor asks a flip side of this question: “I am working on a cure for a rare disease. A wealthy billionaire’s son has the disease. These cuts make it virtually impossible for me to carry on. The billionaire is happy to fund my research but wants all the IP associated with the cure. If I continue my work at the university, the IP and the cure are more likely to be readily available to the public.” So, the question is: are we better off, as a society, with taxpayer funded IP residing with a university or a billionaire?
Room for better cost classifications
The advisor asked me, “my question to you as an accountant would ask much reclassification of costs in the system could take place to make more costs direct, or changing of school/department accounting practices to account for facilities use (is there a best practice based on space utilized or number of personnel?) There hasn’t been a need to be creative up until this point, but I am guessing there could be some work done here (not that there wouldn’t be unintended consequences, I remember my dad stopped having as many grad students and instead used postdocs when he had to start considering them as a direct expense at his school).”
Well, this happens all the time with optimizing billing codes to maximize Medicare and Medicaid reimbursements. Some of this will now creep into areas affected by cuts to overhead reimbursements.
The CEO suggests another unintended consequence of the 15% rate. All else constant, researchers across the system will simply ask for more money to conduct their research. That is unlikely to fly because the new administration has slashed the NIH’s budget from $45 billion to $27 billion making it difficult to support larger grants. But some upcharge is potentially inevitable and will be passed off as the impact of inflation on project costs.
Who will pick up the gap in basic research?
The cuts presume, to some extent, that private donors will step in. But private donors are rarely interested in basic research that pays off over decades not quarters or years. Hence, some shrinkage of basic research is perhaps inevitable.
The question then becomes: will private pharma or laboratories or venture labs now invest more in basic research? Do they have patient capital that will have to wait decades for a payoff to basic research? Will private firms that discover usable basic research extract large rents from other private firms that want to use such research? Is that necessarily socially optimal, given that our prior funding arrangement was that university driven basic research, by and large, created public goods at a highly subsidized rate for the private sector to use? How much of the basic research that was done in US universities simply migrates overseas?
THE FUTURE
Fit for purpose future?
The advisor speculates what the future might look like: “In general, the government has to be the catalyst to support certain areas that don’t have an immediate economic/societal benefit (cancer research, military innovation, nuclear fusion, space travel, etc.), but what organization is best suited might vary.
Research could be done via federal entities, universities, startups, private or public companies, but there will obviously be an immense shortfall in these efforts without the government (maybe the optimistic view is that corporate America becomes the preferred partner for this administration. In certain areas like space exploration, you can already see the money flowing out of NASA and into SpaceX, maybe the for-profit world will capitalize on what will ultimately be a smaller pot of money available).
Ultimately, my gut feeling is that the government should be working to examine the effectiveness of the organizations getting the grant money, in addition to analyzing what lines of grants have had the best results or led to significant progress.”
Or a much scarier future?
The doctor pointed out that significant cuts to overhead reimbursement had been proposed back in the day when Bill Clinton was president and Newt Gingrich was the speaker. That attempt was successfully negated because our leaders back then saw the danger of drastic cuts in funding for the NIH, NSF and similar agencies supporting basic research.
The new cuts are even larger than those proposed in the 1990s. What might happen this time around? The doctor worries that we will get a lost generation of scientists who will stay away from the field given the unpredictability associated with funding research. The brain drain in the field will only help competing nations such as China. As is, we face a large deficit in STEM students relative to China. Chinese STEM PhDs outnumber US counterparts by two to one. If we were to exclude international students from the US numbers, that gap would increase to three to one.
The doctor draws parallels to dark times in history when countries gutted their research programs. For instance, European research potentially never recovered after the Nazi regime caused the exodus of Jewish scientists from that continent to America and elsewhere. Chinese research took a long while to recover from Mao’s purge of academe. The US is the only country in the world that supports basic research in universities at such a massive scale.
Bottomline
The $4 billion savings, reported by NIH, by slashing overhead sounds like a rounding error to me, when we consider our defense budgets or other big ticket spend items such as payments for entitlements such as social security ($ 1.3 trillion in 2023) and Medicare/Medicaid ($ 1.8 trillion). Unfortunately, while the savings may be obvious for now, the intangible costs imposed on healthcare and research in the US are potentially orders of magnitude higher than the $4 billion in savings. Even worse, the impact will start biting a decade from now and may not show up in official statistics in the near future.
Perhaps, this is the new normal. Funding cuts are never easily reversed, regardless of the administration in power. Consider the SALT deduction that was withdrawn, for instance, by the Tax Cuts and Jobs Act (TCJA) in 2017. I am not sure whether it will ever be back in its original form.
In closing, a senior academic leader, who was in a meeting I was at, described the situation well: “after the second world war, the US government struck a grand bargain with US universities to set up the biggest state sponsored venture capital/science experiment apparatus. We give modest sums of money, and you go do the basic research that the country can potentially monetize 10-20 years from now. At universities, we motivated top scientists by paying them say 1/9th of their base salary (an extra month in the summer or say $20-30,000 for folks not used to academic payroll jargon). The best grad students from the world came to work with these top scientists. Enormous payoffs ensued, including the internet, AI and robotics. The ROI to a federal grant dollar to society is easily 2X or 3X or more. And we now run the risk of jeopardizing this amazing innovation and wealth creating machine.”